MELLONI ON THE NOCTURNAL COOLING OF BODIES. 467 



meriting in serene weather, the sky becomes slowly covered with 

 clouds, the differences between the thermometer enveloped in 

 radiating substances and the thermometer with metallic sur- 

 face, diminish very little so long as the clouds remain separated 

 by 30° or 35° from the vertical drawn through the place of ob- 

 servation. As soon as this limit is passed, these differences 

 decrease very rapidly, and totally disappear when the angular 

 space of 35° round the zenith is completely covered w ith clouds. 



But without waiting for these changes in the weathei", which 

 are unfortunately more frequent than the observer occupied with 

 experiments on nocturnal radiation would wish, we may arrive at 

 the same conclusion by a very simple artifice. In fact, it is suf- 

 ficient to incline the vertical axis of the conical vessel which con- 

 tains the bulb of a thermometer with blackened or varnished 

 armature, to perceive that during this period of calm and serenity 

 the thermometer preserves sensibly the same degree of cold until 

 the inclination to the vertical becomes 30° or 35°; beyond this 

 limit the temperature of the thermometer approaches that of the 

 atmosphere, and only differs from it by a small fraction of a 

 degree when the axis of the recipient is brought into the hori- 

 zontal position. 



The law of the proportionality of heat to the sine of the 

 angle formed by the radii with the normal to the element of 

 the radiating surface, led Fourier to the consequence that calo- 

 rific radiation does not proceed solely from the surface of bodies, 

 but also from a certain depth ; a result which followed most 

 clearly from the direct observations of Leslie and Rumford. 



To verify this fact relatively to nocturnal radiation,* prepare 

 two thermometers with corks, so that one may have its armature 

 painted with only a single coat of varnish, the other with eight 

 or ten ; and after having exposed the two instruments to the free 

 air during the night, or still better in their open conical vessels, 

 it will be seen that the first constantly maintains itself higher 

 than the second. At 7 o'clock in the evening on the 19th of 

 September I placed on ray terrace two of these varnished ther- 

 mometers, and a third whose armature was polished. One hour 

 afterwards the three instruments indicated 19°*4, 18°-9, 16°*5; 

 the first indication being that of the metallic surface, the second 

 that of the surface covered with only a single coat of varnish, 

 and the third that of the surface which had ten coats of the same 

 substance superimposed. The thermometers, when observed at 

 9 o'clock, gave 18°'2, l7°-7, 14°'8. We hence conclude that 



2 I 2 



